论文部分内容阅读
解决气液两相流流型的客观、自动识别一直是工业界和学术界所关注的问题。本文以U形管垂直上升段内空气水两相流为例 ,采用压力、压差波动的不同特征参数的组合 ,深入研究了对流型自适应BP网络识别的准确率的影响 ,获得了能够满足工程精度要求的参数波动特征组合方式。本文的结果表明 ,该方法需要建立大量的流型特征的数据库 ,适用于流型的离线识别
Solving the objective and automatic identification of the gas-liquid two-phase flow pattern has always been a concern in industry and academia. Taking the air-water two-phase flow in vertical rising section of U-shaped pipe as an example, this paper uses the combination of different characteristic parameters of pressure and pressure fluctuation to deeply study the influence of convection-based adaptive BP network recognition accuracy, Engineering accuracy requirements of the parameter fluctuation characteristics of the combination. The results of this paper show that this method needs to build a large number of database of flow pattern features and is suitable for offline recognition of flow patterns